[PATCH v3 1/3] spi: rockchip: add support for "cs-gpios" dts property

[PATCH v3 1/3] spi: rockchip: add support for "cs-gpios" dts property

[Jeffy Chen]

Support using "cs-gpios" property to specify cs gpios.

Signed-off-by: Jeffy Chen <jeffy.chen@rock-chips.com>
---

Changes in v3:
include linux/gpio/consumer.h for compile errors on ARCH_X86
(reported by kbuild test robot <lkp@intel.com>)

Changes in v2:
1/ request cs gpios in probe for better error handling
2/ use gpiod* function
(suggested by Heiko Stuebner)
3/ split dt-binding changes to new patch
(suggested by Shawn Lin & Heiko Stuebner)

 drivers/spi/spi-rockchip.c | 31 ++++++++++++++++++++++++++++++-
 1 file changed, 30 insertions(+), 1 deletion(-)

diff --git a/drivers/spi/spi-rockchip.c b/drivers/spi/spi-rockchip.c
index bab9b13..4bcf251 100644
--- a/drivers/spi/spi-rockchip.c
+++ b/drivers/spi/spi-rockchip.c
@@ -16,7 +16,8 @@
 #include <linux/clk.h>
 #include <linux/dmaengine.h>
 #include <linux/module.h>
-#include <linux/of.h>
+#include <linux/gpio/consumer.h>
+#include <linux/of_gpio.h>
 #include <linux/pinctrl/consumer.h>
 #include <linux/platform_device.h>
 #include <linux/spi/spi.h>
@@ -663,6 +664,27 @@ static bool rockchip_spi_can_dma(struct spi_master *master,
 	return (xfer->len > rs->fifo_len);
 }
 
+static int rockchip_spi_setup_cs_gpios(struct device *dev)
+{
+	struct device_node *np = dev->of_node;
+	struct gpio_desc *cs_gpio;
+	int i, nb;
+
+	if (!np)
+		return 0;
+
+	nb = of_gpio_named_count(np, "cs-gpios");
+	for (i = 0; i < nb; i++) {
+		/* We support both GPIO CS and HW CS */
+		cs_gpio = devm_gpiod_get_index_optional(dev, "cs",
+							i, GPIOD_ASIS);
+		if (IS_ERR(cs_gpio))
+			return PTR_ERR(cs_gpio);
+	}
+
+	return 0;
+}
+
 static int rockchip_spi_probe(struct platform_device *pdev)
 {
 	int ret = 0;
@@ -749,6 +771,7 @@ static int rockchip_spi_probe(struct platform_device *pdev)
 	master->transfer_one = rockchip_spi_transfer_one;
 	master->max_transfer_size = rockchip_spi_max_transfer_size;
 	master->handle_err = rockchip_spi_handle_err;
+	master->flags = SPI_MASTER_GPIO_SS;
 
 	rs->dma_tx.ch = dma_request_chan(rs->dev, "tx");
 	if (IS_ERR(rs->dma_tx.ch)) {
@@ -783,6 +806,12 @@ static int rockchip_spi_probe(struct platform_device *pdev)
 		master->dma_rx = rs->dma_rx.ch;
 	}
 
+	ret = rockchip_spi_setup_cs_gpios(&pdev->dev);
+	if (ret) {
+		dev_err(&pdev->dev, "Failed to setup cs gpios\n");
+		goto err_free_dma_rx;
+	}
+
 	ret = devm_spi_register_master(&pdev->dev, master);
 	if (ret) {
 		dev_err(&pdev->dev, "Failed to register master\n");
-- 
2.1.4

[PATCH v3 2/3] dt-bindings: spi/rockchip: add "cs-gpios" optional property

[Jeffy Chen]

Update document devicetree bindings to support "cs-gpios" property.

Signed-off-by: Jeffy Chen <jeffy.chen@rock-chips.com>
---

Changes in v3: None
Changes in v2: None

 Documentation/devicetree/bindings/spi/spi-rockchip.txt | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/Documentation/devicetree/bindings/spi/spi-rockchip.txt b/Documentation/devicetree/bindings/spi/spi-rockchip.txt
index 83da493..d0be2e6 100644
--- a/Documentation/devicetree/bindings/spi/spi-rockchip.txt
+++ b/Documentation/devicetree/bindings/spi/spi-rockchip.txt
@@ -25,6 +25,7 @@ Required Properties:
 
 Optional Properties:
 
+- cs-gpios : Specifies the gpio pins to be used for chipselects.
 - dmas: DMA specifiers for tx and rx dma. See the DMA client binding,
 		Documentation/devicetree/bindings/dma/dma.txt
 - dma-names: DMA request names should include "tx" and "rx" if present.
@@ -48,6 +49,7 @@ Example:
 		#address-cells = <1>;
 		#size-cells = <0>;
 		interrupts = <GIC_SPI 44 IRQ_TYPE_LEVEL_HIGH>;
+		cs-gpios = <&gpio2 23 GPIO_ACTIVE_HIGH>;
 		clocks = <&cru SCLK_SPI0>, <&cru PCLK_SPI0>;
 		clock-names = "spiclk", "apb_pclk";
 		pinctrl-0 = <&spi1_pins>;
-- 
2.1.4

[PATCH v3 3/3] arm64: dts: rockchip: use cs-gpios for cros_ec_spi

[Jeffy Chen]

The cros_ec requires CS line to be active after last message. But the CS
would be toggled when powering off/on rockchip spi, which breaks ec xfer.
Use GPIO CS to prevent that.

Signed-off-by: Jeffy Chen <jeffy.chen@rock-chips.com>
---

Changes in v3: None
Changes in v2:
Fix wrong pinconf for spi5_cs0.

 arch/arm64/boot/dts/rockchip/rk3399-gru.dtsi | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/arch/arm64/boot/dts/rockchip/rk3399-gru.dtsi b/arch/arm64/boot/dts/rockchip/rk3399-gru.dtsi
index eb50593..b34a51d 100644
--- a/arch/arm64/boot/dts/rockchip/rk3399-gru.dtsi
+++ b/arch/arm64/boot/dts/rockchip/rk3399-gru.dtsi
@@ -790,6 +790,8 @@ ap_i2c_audio: &i2c8 {
 &spi5 {
 	status = "okay";
 
+	cs-gpios = <&gpio2 23 GPIO_ACTIVE_HIGH>;
+
 	cros_ec: ec@0 {
 		compatible = "google,cros-ec-spi";
 		reg = <0>;
@@ -813,6 +815,10 @@ ap_i2c_audio: &i2c8 {
 	};
 };
 
+&spi5_cs0 {
+	rockchip,pins = <RK_GPIO2 23 RK_FUNC_GPIO &pcfg_output_high>;
+};
+
 &tsadc {
 	status = "okay";
 
-- 
2.1.4

Re: [PATCH v3 1/3] spi: rockchip: add support for "cs-gpios" dts property

[kbuild test robot]

[-- Attachment #1: Type: text/plain, Size: 1380 bytes --]

Hi Jeffy,

[auto build test ERROR on rockchip/for-next]
[also build test ERROR on v4.12-rc5 next-20170613]
[if your patch is applied to the wrong git tree, please drop us a note to help improve the system]

url:    https://github.com/0day-ci/linux/commits/Jeffy-Chen/spi-rockchip-add-support-for-cs-gpios-dts-property/20170614-160238
base:   https://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip.git for-next
config: x86_64-randconfig-x012-201724 (attached as .config)
compiler: gcc-6 (Debian 6.2.0-3) 6.2.0 20160901
reproduce:
        # save the attached .config to linux build tree
        make ARCH=x86_64 

All errors (new ones prefixed by >>):

   drivers//spi/spi-rockchip.c: In function 'rockchip_spi_probe':
>> drivers//spi/spi-rockchip.c:812:3: error: label 'err_free_dma_rx' used but not defined
      goto err_free_dma_rx;
      ^~~~

vim +/err_free_dma_rx +812 drivers//spi/spi-rockchip.c

   806			master->dma_rx = rs->dma_rx.ch;
   807		}
   808	
   809		ret = rockchip_spi_setup_cs_gpios(&pdev->dev);
   810		if (ret) {
   811			dev_err(&pdev->dev, "Failed to setup cs gpios\n");
 > 812			goto err_free_dma_rx;
   813		}
   814	
   815		ret = devm_spi_register_master(&pdev->dev, master);

---
0-DAY kernel test infrastructure                Open Source Technology Center
https://lists.01.org/pipermail/kbuild-all                   Intel Corporation

[-- Attachment #2: .config.gz --]
[-- Type: application/gzip, Size: 30248 bytes --]

Re: [PATCH v3 1/3] spi: rockchip: add support for "cs-gpios" dts property

[jeffy]

Hi guys(and Robot(^.^)),

this is because i drop this patch:
9783131 New          [v2,1/4] spi: rockchip: fix error handling when probe

which is applied at 6/14, so not on v4.12-rc5 next-20170613.

On 06/14/2017 05:04 PM, kbuild test robot wrote:
> Hi Jeffy,
>
> [auto build test ERROR on rockchip/for-next]
> [also build test ERROR on v4.12-rc5 next-20170613]
> [if your patch is applied to the wrong git tree, please drop us a note to help improve the system]
>
> url:    https://github.com/0day-ci/linux/commits/Jeffy-Chen/spi-rockchip-add-support-for-cs-gpios-dts-property/20170614-160238
> base:   https://git.kernel.org/pub/scm/linux/kernel/git/mmind/linux-rockchip.git for-next
> config: x86_64-randconfig-x012-201724 (attached as .config)
> compiler: gcc-6 (Debian 6.2.0-3) 6.2.0 20160901
> reproduce:
>          # save the attached .config to linux build tree
>          make ARCH=x86_64
>
> All errors (new ones prefixed by >>):
>
>     drivers//spi/spi-rockchip.c: In function 'rockchip_spi_probe':
>>> drivers//spi/spi-rockchip.c:812:3: error: label 'err_free_dma_rx' used but not defined
>        goto err_free_dma_rx;
>        ^~~~
>
> vim +/err_free_dma_rx +812 drivers//spi/spi-rockchip.c
>
>     806			master->dma_rx = rs->dma_rx.ch;
>     807		}
>     808	
>     809		ret = rockchip_spi_setup_cs_gpios(&pdev->dev);
>     810		if (ret) {
>     811			dev_err(&pdev->dev, "Failed to setup cs gpios\n");
>   > 812			goto err_free_dma_rx;
>     813		}
>     814	
>     815		ret = devm_spi_register_master(&pdev->dev, master);
>
> ---
> 0-DAY kernel test infrastructure                Open Source Technology Center
> https://lists.01.org/pipermail/kbuild-all                   Intel Corporation
>

Re: [PATCH v3 1/3] spi: rockchip: add support for "cs-gpios" dts property

[Caesar Wang]

Hi,

As the previous discussed on http://crosreview.com/379681, we have hit a 
failure [0] of ec's xfer, when we support the spi's pd to turn on/off. 
(Says: support the Sdioaudio pd of rk3399 on http://crosreview.com/378562)

[0]:
..
[ 5.579694 ] cros-ec-spi spi5.0: EC failed to respond in time
[ 5.585784 ] cros-ec-spi spi5.0: Transfer error 1/3: -110
..


Feels like that a workaround way to fix it, but that should be a good 
solution.


在 2017年06月14日 11:38, Jeffy Chen 写道:
> Support using "cs-gpios" property to specify cs gpios.
> 
> Signed-off-by: Jeffy Chen <jeffy.chen@rock-chips.com>

Tested-by: Caesar Wang <wxt@rock-chips.com>

I have fetched these patches to test S2R with my board for chromeos4.4.

localhost ~ # cat /sys/kernel/debug/pm_genpd/pm_genpd_summary
...
pd_sdioaudio                    on
     /devices/platform/ff200000.spi                      suspended
     /devices/platform/ff880000.i2s                      active
     /devices/platform/ff8a0000.i2s                      suspended

-Caesar

> ---
> 
> Changes in v3:
> include linux/gpio/consumer.h for compile errors on ARCH_X86
> (reported by kbuild test robot <lkp@intel.com>)
> 
> Changes in v2:
> 1/ request cs gpios in probe for better error handling
> 2/ use gpiod* function
> (suggested by Heiko Stuebner)
> 3/ split dt-binding changes to new patch
> (suggested by Shawn Lin & Heiko Stuebner)
> 
>   drivers/spi/spi-rockchip.c | 31 ++++++++++++++++++++++++++++++-
>   1 file changed, 30 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/spi/spi-rockchip.c b/drivers/spi/spi-rockchip.c
> index bab9b13..4bcf251 100644
> --- a/drivers/spi/spi-rockchip.c
> +++ b/drivers/spi/spi-rockchip.c
> @@ -16,7 +16,8 @@
>   #include <linux/clk.h>
>   #include <linux/dmaengine.h>
>   #include <linux/module.h>
> -#include <linux/of.h>
> +#include <linux/gpio/consumer.h>
> +#include <linux/of_gpio.h>
>   #include <linux/pinctrl/consumer.h>
>   #include <linux/platform_device.h>
>   #include <linux/spi/spi.h>
> @@ -663,6 +664,27 @@ static bool rockchip_spi_can_dma(struct spi_master *master,
>   	return (xfer->len > rs->fifo_len);
>   }
>   
> +static int rockchip_spi_setup_cs_gpios(struct device *dev)
> +{
> +	struct device_node *np = dev->of_node;
> +	struct gpio_desc *cs_gpio;
> +	int i, nb;
> +
> +	if (!np)
> +		return 0;
> +
> +	nb = of_gpio_named_count(np, "cs-gpios");
> +	for (i = 0; i < nb; i++) {
> +		/* We support both GPIO CS and HW CS */
> +		cs_gpio = devm_gpiod_get_index_optional(dev, "cs",
> +							i, GPIOD_ASIS);
> +		if (IS_ERR(cs_gpio))
> +			return PTR_ERR(cs_gpio);
> +	}
> +
> +	return 0;
> +}
> +
>   static int rockchip_spi_probe(struct platform_device *pdev)
>   {
>   	int ret = 0;
> @@ -749,6 +771,7 @@ static int rockchip_spi_probe(struct platform_device *pdev)
>   	master->transfer_one = rockchip_spi_transfer_one;
>   	master->max_transfer_size = rockchip_spi_max_transfer_size;
>   	master->handle_err = rockchip_spi_handle_err;
> +	master->flags = SPI_MASTER_GPIO_SS;
>   
>   	rs->dma_tx.ch = dma_request_chan(rs->dev, "tx");
>   	if (IS_ERR(rs->dma_tx.ch)) {
> @@ -783,6 +806,12 @@ static int rockchip_spi_probe(struct platform_device *pdev)
>   		master->dma_rx = rs->dma_rx.ch;
>   	}
>   
> +	ret = rockchip_spi_setup_cs_gpios(&pdev->dev);
> +	if (ret) {
> +		dev_err(&pdev->dev, "Failed to setup cs gpios\n");
> +		goto err_free_dma_rx;
> +	}
> +
>   	ret = devm_spi_register_master(&pdev->dev, master);
>   	if (ret) {
>   		dev_err(&pdev->dev, "Failed to register master\n");
> 

Re: [PATCH v3 2/3] dt-bindings: spi/rockchip: add "cs-gpios" optional property

[Rob Herring]

On Wed, Jun 14, 2017 at 11:38:02AM +0800, Jeffy Chen wrote:
> Update document devicetree bindings to support "cs-gpios" property.
> 
> Signed-off-by: Jeffy Chen <jeffy.chen@rock-chips.com>
> ---
> 
> Changes in v3: None
> Changes in v2: None
> 
>  Documentation/devicetree/bindings/spi/spi-rockchip.txt | 2 ++
>  1 file changed, 2 insertions(+)

Acked-by: Rob Herring <robh@kernel.org>

Re: [PATCH v3 1/3] spi: rockchip: add support for "cs-gpios" dts property

[jeffy]

Hi doug,

Thanx for your comments.

On 06/23/2017 05:41 AM, Doug Anderson wrote:
> Hi,
>
> On Tue, Jun 13, 2017 at 8:38 PM, Jeffy Chen <jeffy.chen@rock-chips.com> wrote:
>> Support using "cs-gpios" property to specify cs gpios.
>>
>> Signed-off-by: Jeffy Chen <jeffy.chen@rock-chips.com>
>> ---
>>
>> Changes in v3:
>> include linux/gpio/consumer.h for compile errors on ARCH_X86
>> (reported by kbuild test robot <lkp@intel.com>)
>>
>> Changes in v2:
>> 1/ request cs gpios in probe for better error handling
>> 2/ use gpiod* function
>> (suggested by Heiko Stuebner)
>> 3/ split dt-binding changes to new patch
>> (suggested by Shawn Lin & Heiko Stuebner)
>>
>>   drivers/spi/spi-rockchip.c | 31 ++++++++++++++++++++++++++++++-
>>   1 file changed, 30 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/spi/spi-rockchip.c b/drivers/spi/spi-rockchip.c
>> index bab9b13..4bcf251 100644
>> --- a/drivers/spi/spi-rockchip.c
>> +++ b/drivers/spi/spi-rockchip.c
>> @@ -16,7 +16,8 @@
>>   #include <linux/clk.h>
>>   #include <linux/dmaengine.h>
>>   #include <linux/module.h>
>> -#include <linux/of.h>
>> +#include <linux/gpio/consumer.h>
>> +#include <linux/of_gpio.h>
>>   #include <linux/pinctrl/consumer.h>
>>   #include <linux/platform_device.h>
>>   #include <linux/spi/spi.h>
>> @@ -663,6 +664,27 @@ static bool rockchip_spi_can_dma(struct spi_master *master,
>>          return (xfer->len > rs->fifo_len);
>>   }
>>
>> +static int rockchip_spi_setup_cs_gpios(struct device *dev)
>> +{
>> +       struct device_node *np = dev->of_node;
>> +       struct gpio_desc *cs_gpio;
>> +       int i, nb;
>> +
>> +       if (!np)
>> +               return 0;
>
> Not sure you really to check for NULL "np".  Do we really run properly
> without device tree?  We already call of_property_read_u32()
> unconditionally...
hmm, right
>
>
>> +
>> +       nb = of_gpio_named_count(np, "cs-gpios");
>> +       for (i = 0; i < nb; i++) {
>
> Implicitly if there is any error getting "cs-gpios" (AKA if it doesn't
> exist) you'll return a negative value here, then return "0" for the
> function.  AKA cs-gpios is optional...  The behavior is correct, but
> it's a bit non-obvious.  Personally I would have at least put a
> comment even if you didn't put an explicit check.
ok
>
>
>> +               /* We support both GPIO CS and HW CS */
>> +               cs_gpio = devm_gpiod_get_index_optional(dev, "cs",
>> +                                                       i, GPIOD_ASIS);
>> +               if (IS_ERR(cs_gpio))
>> +                       return PTR_ERR(cs_gpio);
>
> As per your discussion with Brian, your whole reason for having this
> function is that:
>
> 1. Core SPI framework treats errors getting the GPIO as non-fatal (SPI
> framework falls back on using the HW chip select).
>
> Mark is the expert, but IMHO that seems like a bug in the core SPI
> framework and you should fix it there rather than hacking around in
> the driver.  _In theory_ you could break backward compatibility
> (someone could have been relying on the old behavior that an error
> caused you to fallback to the HW chip select), but I think that's not
> likely as long as you handle things like:
>
>    cs-gpios = <&gpio1 0 0>, <0>, <&gpio1 1 0>, <&gpio1 2 0>;
>
> AKA if someone has explicitly specified <0> for the GPIO then _that_
> shouldn't be an error and we should do the fallback to HW chip select.
> If we really expect old buggy DTS files that get broken by the old
> behavior then we'd have to ask for advice from Mark and/or device tree
> experts...
right, it would be good to be handled in the spi core.
and i think devm_gpiod_get_index_optional would take care of the <0> 
fallback case
>
>
> As evidence that the current SPI core is broken in the way it is and
> could use a patch, it would be easy to "fall back" to a chip select
> that's greater than "master->num_chipselect", which seems to me like a
> clear bug.
>
> --
>
> 2. The SPI framework doesn't end up calling gpiod_request().  It seems
> like it ought to.  Requiring the sub driver to do this seems wrong.
>
>
>> +       }
>> +
>> +       return 0;
>> +}
>> +
>>   static int rockchip_spi_probe(struct platform_device *pdev)
>>   {
>>          int ret = 0;
>> @@ -749,6 +771,7 @@ static int rockchip_spi_probe(struct platform_device *pdev)
>>          master->transfer_one = rockchip_spi_transfer_one;
>>          master->max_transfer_size = rockchip_spi_max_transfer_size;
>>          master->handle_err = rockchip_spi_handle_err;
>> +       master->flags = SPI_MASTER_GPIO_SS;
>
> IMHO this one line in your patch makes total sense and it seems like
> you could post it by itself and it could land.  All the error check
> and gpiod_request() bikeshedding could be deferred to a separate
> patch.
ok, that make sense, will do in next version.
>
>
>